Hands-On Python Natural Language Processing: Explore tools and techniques to analyze and process text with a view to building real-world NLP applications. Code

Cover
Title Page
Copyright and Credits
About Packt
Contributors
Table of Contents
Preface
Section 1: Introduction
Chapter 1: Understanding the Basics of NLP
	Programming languages versus natural languages
		Understanding NLP
	Why should I learn NLP?
	Current applications of NLP
		Chatbots
		Sentiment analysis
		Machine translation
		Named-entity recognition
		Future applications of NLP
	Summary
Chapter 2: NLP Using Python
	Technical requirements
	Understanding Python with NLP 
		Python\'s utility in NLP
	Important Python libraries
		NLTK
			NLTK corpora
				Text processing
				Part of speech tagging
		Textblob
			Sentiment analysis
			Machine translation
			Part of speech tagging
		VADER
	Web scraping libraries and methodology
	Overview of Jupyter Notebook
	Summary
Section 2: Natural Language Representation and Mathematics
Chapter 3: Building Your NLP Vocabulary
	Technical requirements
	Lexicons
	Phonemes, graphemes, and morphemes
	Tokenization 
		Issues with tokenization
		Different types of tokenizers
			Regular expressions 
			Regular expressions-based tokenizers
			Treebank tokenizer
			TweetTokenizer 
	Understanding word normalization
		Stemming
			Over-stemming and under-stemming
		Lemmatization 
			WordNet lemmatizer
			Spacy lemmatizer
		Stopword removal
		Case folding
		N-grams
		Taking care of HTML tags
		How does all this fit into my NLP pipeline?
	Summary
Chapter 4: Transforming Text into Data Structures
	Technical requirements
	Understanding vectors and matrices
		Vectors
		Matrices
	Exploring the Bag-of-Words architecture
		Understanding a basic CountVectorizer
		Out-of-the-box features offered by CountVectorizer
			Prebuilt dictionary and support for n-grams
			max_features
			Min_df and Max_df thresholds
		Limitations of the BoW representation
	TF-IDF vectors
		Building a basic TF-IDF vectorizer
		N-grams and maximum features in the TF-IDF vectorizer 
		Limitations of the TF-IDF vectorizer\'s representation
	Distance/similarity calculation between document vectors
		Cosine similarity
			Solving Cosine math
			Cosine similarity on vectors developed using CountVectorizer
			Cosine similarity on vectors developed using TfIdfVectorizers tool
	One-hot vectorization
	Building a basic chatbot
	Summary 
Chapter 5: Word Embeddings and Distance Measurements for Text
	Technical requirements
	Understanding word embeddings
	Demystifying Word2vec
		Supervised and unsupervised learning
		Word2vec – supervised or unsupervised?
		Pretrained Word2vec 
		Exploring the pretrained Word2vec model using gensim
		The Word2vec architecture
			The Skip-gram method
				How do you define target and context words?
			Exploring the components of a Skip-gram model
				Input vector
				Embedding matrix
				Context matrix
				Output vector
				Softmax
				Loss calculation and backpropagation
				Inference
			The CBOW method
			Computational limitations of the methods discussed and how to overcome them
				Subsampling
				Negative sampling
				How to select negative samples
	Training a Word2vec model 
		Building a basic Word2vec model
		Modifying the min_count parameter 
		Playing with the vector size
		Other important configurable parameters
		Limitations of Word2vec
		Applications of the Word2vec model 
	Word mover’s distance
	Summary
Chapter 6: Exploring Sentence-, Document-, and Character-Level Embeddings
	Technical requirements
	Venturing into Doc2Vec
		Building a Doc2Vec model
			Changing vector size and min_count 
			The dm parameter for switching between modeling approaches
			The dm_concat parameter
			The dm_mean parameter
			Window size
			Learning rate
	Exploring fastText 
		Building a fastText model
		Building a spelling corrector/word suggestion module using fastText
		fastText and document distances
	Understanding Sent2Vec and the Universal Sentence Encoder
		Sent2Vec
		The Universal Sentence Encoder
	Summary 
Section 3: NLP and Learning
Chapter 7: Identifying Patterns in Text Using Machine Learning
	Technical requirements
	Introduction to ML
	Data preprocessing
		NaN values
		Label encoding and one-hot encoding
		Data standardization
			Min-max standardization
			Z-score standardization
	The Naive Bayes algorithm
		Building a sentiment analyzer using the Naive Bayes algorithm
	The SVM algorithm
		Building a sentiment analyzer using SVM
	Productionizing a trained sentiment analyzer
	Summary 
Chapter 8: From Human Neurons to Artificial Neurons for Understanding Text
	Technical requirements
	Exploring the biology behind neural networks
		Neurons
		Activation functions
			Sigmoid
			Tanh activation
			Rectified linear unit
		Layers in an ANN
	How does a neural network learn?
		How does the network get better at making predictions?
	Understanding regularization
		Dropout
	Let\'s talk Keras
	Building a question classifier using neural networks
	Summary
Chapter 9: Applying Convolutions to Text
	Technical requirements
	What is a CNN?
		Understanding convolutions
			Let\'s pad our data
			Understanding strides in a CNN
		What is pooling?
		The fully connected layer
	Detecting sarcasm in text using CNNs
		Loading the libraries and the dataset
		Performing basic data analysis and preprocessing our data
		Loading the Word2Vec model and vectorizing our data
		Splitting our dataset into train and test sets
		Building the model
		Evaluating and saving our model
	Summary
Chapter 10: Capturing Temporal Relationships in Text
	Technical requirements
	Baby steps toward understanding RNNs
		Forward propagation in an RNN
		Backpropagation through time in an RNN
	Vanishing and exploding gradients
	Architectural forms of RNNs
		Different flavors of RNN
		Carrying relationships both ways using bidirectional RNNs
		Going deep with RNNs
	Giving memory to our networks – LSTMs
		Understanding an LSTM cell
			Forget gate
			Input gate
			Output gate
		Backpropagation through time in LSTMs
	Building a text generator using LSTMs
	Exploring memory-based variants of the RNN architecture
		GRUs
		Stacked LSTMs
	Summary
Chapter 11: State of the Art in NLP
	Technical requirements
	Seq2Seq modeling
		Encoders
		Decoders
			The training phase
			The inference phase
	Translating between languages using Seq2Seq modeling 
	Let\'s pay some attention
	Transformers 
		Understanding the architecture of Transformers
			Encoders 
			Decoders
			Self-attention
				How does self-attention work mathematically?
				A small note on masked self-attention
			Feedforward neural networks
			Residuals and layer normalization
			Positional embeddings
			How the decoder works
			The linear layer and the softmax function
			Transformer model summary
	BERT 
		The BERT architecture
		The BERT model input and output
		How did BERT the pre-training happen?
			The masked language model
			Next-sentence prediction 
		BERT fine-tuning
	Summary
Other Books You May Enjoy
Index